Total No. of Questions : 6] SEAT No.

:
P4209 [4760] - 1191 [Total No. of Pages :3

M.E. (Computer)
ADVANCED COMPUTER ARCHITECTURE
(2013 Credit Course) (510103) (Semester - I)
Time : 3Hours] [Max. Marks : 50
Instructions to the candidates:
1) All Questions are compulsory.
2) Neat diagrams must be drawn wherever necessary.
3) Figures to the right side indicate full marks.
4) Assume Suitable data if necessary.

Q1) a) Compare with features different shared memory multiprocessor

models. [5]
b) State the following terms w.r.t. various interconnect architectures: [4]
i) Network diameter
ii) Bisection Bandwidth
iii) Routing function
iv) Multistage Network
OR
a) Draw the dependence graph and analyze the various dependencies among
the following statements in a given program. [5]
S1 : Load R1, M(100)
S2 : Move R2, R1
S3 : Inc R1
S4 : ADD R2, R1
S5 : Store M(100), R1
b) State and compare Interconnection Networks used in multiprocessor
systems. [4]

P.T.O.
Q2) a) Define the term Degree of Parallelism (DOP). Describe Average Parallelism
in terms of DOP. [4]
b) With example, illustrate the mismatch between software and hardware
parallelism. How can we match the software and hardware parallelism?[4]
OR
a) What is Scalability? Define the term Speedup and Efficiency w.r.t.
scalability? How these parameters reflects the performance of parallel
computer systems? [4]
b) Derive the Amdahl’s law for speedup performance. Comment on the
major observations and conclusions drawn w.r.t. the speedup obtained.[4]

Q3) a) Compare the features and performance of CISC and RISC processor
architectures. [4]
b) How arithmetic and instruction pipelines are designed? What is the use
of Reservation Table? [4]
OR
a) State the 4-level memory hierarchy defined for a computer system. How
the data transfer takes place between adjacent levels of a memory
hierarchy? [4]
b) Discuss in brief the different mechanisms implemented in superscalar
processor architectures to enhance the performance of Instruction and
Arithmetic pipeline. [4]

Q4) a) With example state the advantages of vector processing over scalar
processing. What is Vectorizing Compiler? State any 2 Vector optimizing
functions. [4]
b) What is cache coherency? What are the cache write policies used for
cache updating? State the 4 states of MESI protocol. [4]
OR
a) Discuss and compare between store-and-forward routing and wormhole
message routing schemes. [4]
b) With example explain the use of Compound Vector Function (CVF) to
perform the vector operations. [4]

[4760]-1191 2
Q5) a) With example explain message passing parallel programming. What is
SPMD Programming? [4]
b) Explain with example the use of synchronization primitives in parallel
programming. [4]
OR
a) State and discuss different collective communication functions designed
for MPI. [4]
b) Explain the support provided by an OS for parallel program execution.
Comment on thread and process level parallelism. [4]

Q6) a) Explain different services offered by Cloud. What is the difference

between public and private cloud? [5]
b) Discuss important features of Quantum computing. How these
architectures can be used for distributed parallel processing? [4]
OR
a) Compare between grid and cloud computing. What is cloud
middleware? [5]
b) How Neural Networks can be used for distributed parallel computing?
Discuss in brief. [4]

E E E

[4760]-1191 3